# Geometric MA

A Trend reference entry (Moving-average lineage) in the LuxAlgo Library: explained, not implemented as a chart indicator.

## What is a Geometric MA?

A geometric MA replaces the arithmetic mean inside a simple moving average with the geometric mean: multiply the last N prices together and take the Nth root. The equivalent, numerically safer construction is to average the logarithms of price and exponentiate the result, which makes the geometric MA an [SMA](https://www.luxalgo.com/library/concept/sma/) computed in log space. By the AM-GM inequality it always sits at or below the arithmetic average of the same window, and the gap widens as dispersion inside the window grows.

Geometric averaging has a longer pedigree in markets than the chart overlay does. The Financial Times' FT 30 index, launched in 1935, combined its constituents with a geometric mean, and the Value Line Geometric Composite has averaged a broad equity list the same way since 1961; both choices reflect the same multiplicative logic, and both are known to drift below an arithmetic equivalent over time. The moving-average version imports that logic to a rolling window.

The rationale is compounding: returns multiply rather than add, so averaging in log space matches how prices actually grow and pairs naturally with logarithmic chart scales. On short windows over typical price series the geometric and arithmetic lines are nearly indistinguishable; the distinction earns its keep on long horizons, volatile series, and percentage-based analysis.

In practice the line behaves like its arithmetic sibling: same window, same lag, nearly the same path, sitting at or slightly below it. Two constraints are worth knowing. The construction needs strictly positive inputs, since logarithms of zero or negative values are undefined, so it suits prices but not oscillators or spread series that cross zero. And because the difference from the arithmetic line is a level shift rather than a timing change, switching mean types will not rescue a late crossover system.

## How to identify a geometric MA on a chart

Visually a geometric MA is almost a twin of the arithmetic average, so identification relies on comparison and construction rather than shape.

1. Plot the geometric MA against a same-length simple moving average: the geometric line sits at or just below it everywhere, never above.
2. Watch volatile stretches: the gap between the two lines widens as dispersion inside the window rises and closes again in calm conditions.
3. Switch to logarithmic scale: the geometric MA is a plain arithmetic average in log space, so its behavior looks most natural there.
4. Inspect the indicator's source if available: the fingerprints are an average of log prices followed by an exponential, or an N-term product followed by an Nth root.
5. Confirm the input domain: a genuine geometric MA fails or misbehaves on series containing zeros or negative values, which an arithmetic average tolerates.

## How it's calculated

The geometric moving average is the n-th root of the product of the last n prices, an average that treats percentage changes symmetrically.

```
GMA_t = (P_t × P_{t-1} × ... × P_{t-n+1})^(1/n)
Equivalent log form: GMA_t = exp((Σ ln(P_{t-i}) over i = 0..n-1) / n)

  P_t: source price at bar t (typically the close)
  t: bar index
  n: lookback length in bars (commonly 14)
  i: offset inside the window
  ln: natural logarithm
  exp: natural exponential function, e raised to the argument
  GMA_t: geometric moving average at bar t
```

Equals the exponential of a simple moving average of log prices, which is how it is usually computed to avoid overflow on long windows.

It sits at or below the arithmetic SMA of the same window (AM-GM inequality), with the gap widening as volatility rises.

Inputs must be strictly positive, so it is applied to prices, not to returns or oscillators that can hit zero or go negative.

## How traders use it

- As a trend baseline on log-scaled or long-horizon charts, where log-space averaging keeps the line consistent with percentage moves rather than point moves.
- In compounded-growth work: geometric averaging underlies compound-return statistics, so geometric smoothing shows up in long-cycle growth bands and valuation-style curves.
- As a dispersion tell: since the shortfall below the SMA scales with variance, the spread between same-length arithmetic and geometric averages is a crude volatility read.
- In crossover systems: geometric fast and slow lines, or a geometric baseline against an [EMA](https://www.luxalgo.com/library/concept/ema/), run through the same [moving average crossover](https://www.luxalgo.com/library/concept/moving-average-crossovers/) mechanics as arithmetic versions, with signals differing only marginally on most series.
- As long-horizon structure: on multi-year log-scale charts a geometric baseline serves as [dynamic support and resistance](https://www.luxalgo.com/library/concept/dynamic-s-r-via-ma/), and its direction can be read through an [MA slope filter](https://www.luxalgo.com/library/concept/ma-slope-filter/) or stacked with other lengths into an [MA ribbon](https://www.luxalgo.com/library/concept/ma-ribbon/).
- As an envelope base: a percentage-width [MA envelope](https://www.luxalgo.com/library/concept/ma-envelope/) around a geometric baseline is internally consistent, since both the center line and the offsets are multiplicative objects.

## Geometric MA vs other moving averages

- **SMA** (https://www.luxalgo.com/library/concept/sma/): The only difference is the mean: arithmetic versus geometric over the same window. By the AM-GM inequality the geometric line is never higher, and the shortfall grows with the window's variance; the timing of turns and crossovers is essentially unchanged.
- **EMA** (https://www.luxalgo.com/library/concept/ema/): The EMA changes the weighting, front-loading recent bars to cut lag, while a geometric MA changes the type of mean and keeps equal weights. They answer different questions, and the two ideas can be combined by running exponential weights on log prices.
- **VWMA** (https://www.luxalgo.com/library/concept/vwma/): VWMA reweights the arithmetic mean by volume so heavy-turnover bars dominate. A geometric MA keeps equal weights but averages multiplicatively; one adjusts for participation, the other for compounding.

## FAQ

### How different is a geometric moving average from an SMA?

On short, calm windows the two are nearly identical. The geometric version is always equal or slightly lower, and in log terms the shortfall is approximately half the variance of the window's log values, so the gap widens as volatility rises. For day-to-day smoothing the visual difference is negligible; it becomes meaningful on long windows, volatile series, and percentage-based work.

### Why use a geometric mean for prices at all?

Because prices compound. A series that gains 10% and then loses 10% does not finish flat, and arithmetic averaging quietly ignores that asymmetry. The geometric mean is the average consistent with multiplicative growth, which is why compound-return statistics are built on it. As a plotted smoother the effect is subtle, but for return-based or log-scale analysis it is the coherent choice.

### Is a geometric moving average faster or less laggy than an SMA?

No. Lag comes from the window length and the weighting, and a geometric MA keeps both identical to the simple average; only the level changes, sitting slightly lower by an amount tied to the window's variance. Traders wanting less lag should look at exponential or adaptive weightings rather than a different type of mean.

### Can a geometric MA be used on oscillators or spreads?

Not safely. The geometric mean requires strictly positive inputs because it multiplies values or averages their logarithms, and a single zero or negative reading breaks the computation. It belongs on price series and other strictly positive quantities; for zero-crossing series, stick with arithmetic averaging.

### Which indicators implement geometric moving averages?

Implementations appear under several names: dedicated log-space averages such as RezzaHmt's Logarithmic Moving Average, and generalized power-mean tools like BackQuant's Step Generalized Moving Average, where the geometric mean is the special case as the exponent approaches zero. LuxAlgo's Geometric Bias Oscillator draws on the same geometric-mean arithmetic to build a directional-bias reading.

### Should the chart be on log scale when using a geometric MA?

It is the natural pairing, since the geometric MA is an arithmetic average in log space and a log-scaled chart renders equal percentage moves as equal distances. Nothing breaks on a linear chart, and on short windows the choice is cosmetic, but for multi-year analysis the combination keeps the average and the axis telling the same percentage-based story.

## Related concepts

- SMA: https://www.luxalgo.com/library/concept/sma/
- EMA: https://www.luxalgo.com/library/concept/ema/
- Adaptive-lookback MA: https://www.luxalgo.com/library/concept/adaptive-lookback-ma/
- MA Envelope: https://www.luxalgo.com/library/concept/ma-envelope/
- SWMA: https://www.luxalgo.com/library/concept/swma/
- RMA: https://www.luxalgo.com/library/concept/rma/
- HMA: https://www.luxalgo.com/library/concept/hma/
- KAMA: https://www.luxalgo.com/library/concept/kama/
- JMA: https://www.luxalgo.com/library/concept/jma/
- ZLEMA: https://www.luxalgo.com/library/concept/zlema/

---

Source: https://www.luxalgo.com/library/concept/geometric-ma/ (LuxAlgo Library, the encyclopedia of trading & technical analysis). Free to use with attribution: https://www.luxalgo.com/library/license/